J Ornithol
DOI 10.1007/s10336-014-1088-y
ORIGINAL ARTICLE
Aggression, body condition, and seasonal changes in sex-steroids
in four hummingbird species
Paulina L. González-Gómez • William S. Blakeslee
Pablo Razeto-Barry • Rachel M. Borthwell •
Sara M. Hiebert • John C. Wingfield
Received: 24 September 2013 / Revised: 27 April 2014 / Accepted: 22 May 2014
Ó Dt. Ornithologen-Gesellschaft e.V. 2014
Abstract Hummingbirds present a unique combination
between extremely high life costs and a number of efficient
adaptations to fuel these demands. In addition to cognitive
abilities, territorial hummingbirds display aggressive
behaviors that allow for access to better food resources. In
year-round territorial species, male–male territorial
aggression is similar between breeding and non-breeding
seasons; however, the endocrine mechanisms underlying
control of territoriality during these distinct seasonal peri-
ods may differ. In many species, testosterone (T) triggers
increased aggression during the breeding season whereas
territoriality in the non-breeding season can be regulated by
circulating the biologically inert sex steroid precursor
dehydroepiandrosterone (DHEA) and converting it to T in
target tissues. The seasonal hormonal regulation of hum-
mingbird territorial behavior has heretofore been unknown.
Our goal was to assess seasonal changes in sex steroids,
territorial aggression levels, and body condition during
reproductive and non-reproductive seasons in
P. L. González-Gómez  W. S. Blakeslee 
R. M. Borthwell  J. C. Wingfield
Department of Neurobiology, Physiology, and Behavior,
University of California Davis, 1 Shields Avenue, Davis,
CA 95616, USA
P. L. González-Gómez (&)  P. Razeto-Barry
Instituto de Filosofı́a y Ciencias de la Complejidad (IFICC),
Los Alerces, 3024 Santiago, Chile
e-mail: plgonzalezgomez@gmail.com
P. Razeto-Barry
Universidad Diego Portales, Vicerrectorı́a Académica,
Manuel Rodrı́guez Sur 415, Santiago, Chile
S. M. Hiebert
Department of Biology, Swarthmore College,
500 College Avenue, Swarthmore, PA 19081, USA
•
hummingbirds. To validate the use of cloacal fluid (CF) for
the study of sex steroids, steroid levels in plasma and CF
were correlated in Sephanoides sephaniodes. During the
reproductive season, Calypte. anna, Archilochus alexandri,
and Selasphorus rufus males showed high levels of T that
were positively correlated with aggression, but the rela-
tionship between T and body condition was not consistent
across species. As expected, T levels in females were
significantly lower than in males in all seasons, however
still detectable. During the non-reproductive season, CF
DHEA of Calypte anna was high and positively correlated
with aggressive behaviors and body condition. Our results
suggest that hummingbirds display aggressive behaviors
that could be linked to different hormones during the
breeding and non-breeding seasons.
Keywords Testosterone  Dehydroepiandrosterone 
Hummingbirds  Territoriality  Body condition 
Seasonality  Aggression
Zusammenfassung
Aggressionsverhalten, Körperkondition und jahreszei-
tliche Schwankungen der Sexualsteroidspiegel bei vier
Kolibriarten
Kolibris verbinden in einzigartiger Weise eine extrem
kostspielige Lebensweise mit einer Reihe von Anpas-
sungsleistungen, um ihren Energiebedarf effizient zu dec-
ken. Zusätzlich zu ihren kognitiven Fähigkeiten verfügen
territoriale Kolibris über aggressive Verhaltensweisen, die
ihnen den Zugang zu besseren Nahrungsquellen sichern.
Bei Arten, die ganzjährig Reviere verteidigen, ist die ter-
ritoriale Aggression zwischen Männchen zur Brutzeit
ähnlich wie außerhalb; allerdings können sich die der
123

Steuerung der Territorialität zugrunde liegenden endokri-
nen Mechanismen während dieser jahreszeitlich klar ge-
trennten Zeiträume unterscheiden. Bei vielen Arten löst
Testosteron (T) eine Zunahme des Aggressionsverhaltens
während der Brutsaison aus, wohingegen die Territorialität
außerhalb der Brutzeit durch den Einsatz des biologisch
inerten Sexualsteroid-Vorläufers Dehydroepiandrosteron
(DHEA) reguliert werden kann, welcher dann im Ziel-
gewebe in T umgewandelt wird. Die jahreszeitliche hor-
monelle Steuerung des Territorialverhaltens bei Kolibris
war bislang unbekannt. Unser Ziel war es, jahreszeitliche
Schwankungen des Spiegels von Sexualsteroiden, das
Ausmaß territorialer Aggression sowie die Körperkondi-
tion der Kolibris jeweils während und außerhalb der Fort-
pflanzungszeit zu ermitteln. Um die Eignung der
Kloakenflüssigkeit (cloacal fluid, CF) für die Untersuchung
von Sexualsteroiden zu testen, wurden bei Sephanoides
sephaniodes die jeweiligen Steroidspiegel in Plasma und
CF miteinander in Bezug gesetzt. Während der Fort-
pflanzungsperiode zeigten Männchen von C. anna, Ar-
chilochus alexandri und Selasphorus rufus jeweils hohe
T-Spiegel, welche positiv mit dem Aggressionsverhalten
korrelierten; das Verhältnis zwischen T und der Körper-
kondition stimmte jedoch nicht bei allen Arten überein.
Erwartungsgemäß lagen die T-Spiegel bei den Weibchen
zu allen Jahreszeiten signifikant niedriger als bei den
Männchen, waren aber trotzdem nachweislich vorhanden.
Außerhalb der Fortpflanzungszeit lag der CF-Wert für
DHEA bei Calypte anna hoch und korrelierte positiv mit
dem Aggressionsverhalten und der Körperkondition. Un-
sere Ergebnisse legen nahe, dass die von Kolibris gezeigten
aggressiven Verhaltensweisen zur Brutzeit beziehungswe-
ise außerhalb dieser an verschiedene Hormone gekoppelt
sein könnten.
Introduction
Energy balance in hummingbirds is the result of a complex
interplay between foraging behavior, physiological con-
straints, cognitive performance, and environmental condi-
tions (Hainsworth 1978; Suarez and Gass 2002). During
hovering, hummingbirds achieve some of the highest mass-
specific rates of aerobic metabolism known among verte-
brates (Suarez 1992). Moreover, their small body size
imposes high thermoregulatory costs (González-Gómez
et al. 2011a). In order to fuel these high metabolic rates,
hummingbirds are able to remember the location and
availability of the best resources (González-Gómez et al.
2011b). In fact, recent studies suggest individuals with
better cognitive abilities exhibit higher energy intake
(González-Gomez et al. 2011c).
123
J Ornithol
In some hummingbirds species such as the Green-backed
Firecrown hummingbird (Sephanoides sephaniodes) and
the Annas hummingbird (Calypte anna), males display two
strategies, territorial individuals, which actively defend
feeding resources by displaying aggressive behaviors, such
as performing acrobatic aerial displays, vocalizing, and
chasing intruders and non-territorial individuals, which are
less aggressive and mostly opportunistic (González-Gómez
et al. 2011a; González-Gómez, non-published data). Pre-
vious studies conducted by González-Gómez et al. (2011a)
showed that, when the cost of aggression is higher than the
benefit derived from the defended nectar sources, territorial
behaviors cease. Therefore, it is hypothesized that territorial
males display a positive correlation between aggression and
body condition. Females, in contrast, are in general non-
territorial and either perform rapid intrusions into male
territories to raid high-quality resources (González-Gómez
and Estades 2009) or are forced to forage in nectar-poor,
scattered resources (Kodric-Brown and Brown 1978;
Temeles et al. 2005).
Even though the behavioral patterns of male–male ter-
ritorial aggression are similar between breeding and non-
breeding seasons in year-round territorial avian species,
there is extensive evidence showing that endocrine regu-
lation mechanisms are different both seasons in most of
bird species (Hau et al. 2000; Wingfield et al. 2001;
Wingfield and Soma 2002; Hau 2007; Goymann et al.
2007). In the breeding season, it is well-known that tes-
tosterone (T) triggers an increase in male–male competi-
tion for territories and mates (Wingfield 1984; Ogawa et al.
1997). Possibly because high levels of testosterone could
induce inappropriate behaviors outside of the breeding
season such as sexual behavior (Hau 2007), and could have
detrimental effects such as decreasing body mass and fat
stores (Wingfield and Soma 2002), alternatives to T have
been proposed outside of the breeding season. One possi-
bility could be dehydroepiandrosterone (DHEA) a biolog-
ically inert androgen precursor that can be converted to
biologically active sex steroids in the brain, thus, avoiding
peripheral androgenization in a non-reproductive context
(Soma and Wingfield 2001; Soma et al. 2000, 2002). For
example, simulated territorial intrusions performed with
spotted antbirds during the non-reproductive season (i.e.,
birds had regressed gonads and low levels of androgen and
estrogen) showed that plasma DHEA was directly corre-
lated to the intensity of aggressive displays in both males
and females (Hau et al. 2000). Interestingly, hormonal
regulation of hummingbird territorial behavior has hereto-
fore been unstudied. In this study, we aimed to assess
seasonal changes in sex steroids, territorial aggression
levels, and body condition during the reproductive and
non-reproductive seasons in four hummingbird species that
defend food resources such as feeders and flowering trees
J Ornithol
and shrubs. In this context, we expected that (a) males
would have higher levels of testosterone than females
during the reproductive season; (b) higher levels of
aggression in territorial males would be correlated with
higher levels of T during the reproductive season and
higher levels of DHEA during the non-reproductive season;
and (c) body condition would be positively related to
DHEA or T levels in males.
Methods
To assess DHEA and T levels in reproductive and non-
reproductive seasons and to determine the relationship
between hormone levels in plasma and CF samples, we
performed a laboratory study in Chile. Field studies were
conducted in California, USA to evaluate the relationship
between hormones in CF, aggressive behaviors, and body
condition. All protocols were conducted according to
Chilean laws, legal permits and the ethical committees of
the Universities (IACUC Protocol No. 16095, Federal Bird
Banding Permit No. 21535).
Laboratory study
The study was carried out during April (non-reproductive
season, n = 5) and July (beginning of the reproductive
season, n = 4) of 2009 in the Andean foothills at the
Mediterranean field station of the Universidad Católica de
Chile at San Carlos de Apoquindo, central Chile (33°230 S,
70°310 W, 1,100 m elevation above sea level). Subjects
were male Green-backed Firecrown hummingbirds, Sep-
hanoides sephaniodes (5.8 ± 0.19 g) that had established
and were actively defending feeding territories at the study
site. Birds were caught with mist nets and housed for seven
days in individual mesh cages (1 9 1 9 1 m) under nat-
ural photoperiod and temperature at the field site. Birds
were fed with nutritionally complete, artificial nectar
(Nektar Plus, Nekton USA, Clearwater, FL). All birds were
released after the study.
Field studies
Observations and data collection were carried out from
September 2010 to August 2011 during the reproductive
and non-reproductive seasons, with the objective of
relating aggression, hormone levels, and body condition.
The study was carried out in Chalk Hill and Baggins
End, Davis, CA. Subjects were free-living Anna’s
hummingbirds (Calypte anna, 4.82 ± 0.1 g), which were
sampled from September to early December (non-
reproductive season, n = 9 males) and from middle
January to middle May (reproductive season, n = 13
females and 34 males); Rufous hummingbirds (Selas-
phorus rufus, 3.15 ± 0.05 g), which were sampled from
March to April (migratory/reproductive season, n = 9
females and 6 males; see below); and Black-chinned
hummingbirds (Archilochus alexandri, 3.3 ± 0.06 g),
which were sampled from April to late June (repro-
ductive season, n = 15 females and 7 males) and from
middle August to early September (non-reproductive
season, n = 8 females and 10 males).
In order to differentiate properly between territorial and
non-territorial hummingbirds in a free-living environment,
a training period was conducted. Feeders were introduced
into the surrounding environment and observations made
every day between 8:00 and 11:30 a.m. for a week. Sub-
jects that actively and consistently defended the feeders as
part of their territories for more than a day were classified
as ‘‘territorial birds’’. Individuals that were able to use the
feeders and occasionally displayed aggressive behaviors,
but were not consistently observed defending or in the
feeder’s surrounding area, were classified as ‘‘non-territo-
rial’’ individuals. The artificial feeder consisted of a com-
mercial 50 ml syringe enveloped with red paper. Each
feeder was filled twice a day with 50 ml of 25 % w/w
sucrose solution. On the day of hormone sampling, the
feeder was set up in a cylindrical trap (30 cm diameter,
50 cm height) and aggressive territorial behaviors were
observed for 1 h. All the observations were performed
between 7:00 and 11:00 a.m. to avoid behavioral variability
throughout the day. To assess the intensity of territorial
defense, an aggressiveness index described by González-
Gómez et al. (2011a) was utilized. In short, this index is
composed of three different territorial behaviors: aggres-
sive calls (vocalizations, V), chases (C), contacts (aggres-
sion, O), and their combinations weighted by the energy
used in the aggressive action. We obtained that V/C/O has
the relations 2:3:5, and therefore, we constructed the fol-
lowing aggressive index (AI):
AI ¼ V  2 þ C  5 þ VC  7 þ CO  8 þ VCO  10:
Birds were captured in the trap immediately after the
period of observation and CF was collected. All CF sam-
ples were collected in the first 3 min after capture. In
addition, body weight and tarsus length were recorded for
each bird. Birds were weighed with a Pesola Micro-Line
Spring Scale (capacity 10 g) and measured with a Fowler
(Brantford, Ontario) UltraCal Mark III Electronic Caliper.
Body condition index was calculated as the residuals of the
regression between tarsus length and body mass
(R2 = 0.17, p = 0.03). Because of the volume constraint
(i.e., each sample was roughly 30–50 lL), CF samples
were analyzed for either T (reproductive season, Anna’s,
Black-chinned and Rufus hummingbirds) or DHEA (non-
123

reproductive season, Anna’s hummingbirds) using a
radioimmunoassay (RIA) as described below.
Blood samples
A 30 lL blood sample (roughly 5 % of body mass) was
taken from the tarsus vein immediately after capture with a
heparinized capillary tube. Tubes were sealed using sealing
clay. This procedure was repeated 72 h later in order to
obtain the necessary volume to perform the hormone ana-
lysis (i.e., 25 lL of plasma). All birds maintained or
increased body mass during captivity and were released
after two days of the last blood collection when we were
confidant they had recovered. Samples were immediately
centrifuged at 2,000 rpm for 5 min and plasma was moved
to a -20 °C freezer.
Cloacal fluid samples
Samples were collected non-invasively by holding an 80
lL microcapillary tube directly outside the cloaca. During
aspiration, care was taken to avoid fecal masses. If a
sample had visible fecal masses, it was discarded. Samples
were frozen immediately after collection in a -20 °C
freezer.
Controlling for urine dilution
In order to control for the animal’s hydration state and its
possible effect on hormone levels, CF creatinine was used
as a reference for S. sephaniodes (Hiebert et al. 2000).
Creatinine concentration was determined with reagents
from Cayman Chemical (Cat No. 500701), following the
protocol described by Hiebert et al. (2000).
The levels of creatinine were undetectable in CF from the
other species in this study. Instead, specific gravity (here-
after SG) was used, which refers to the density (mass per
volume) of a sample in comparison to pure water (Arm-
strong 2005). SG was determined using a clinical temper-
ature compensated refractometer (Westover Scientific,
RHC200ATC). Unfortunately, due to the limited amount of
CF obtained from free-living hummingbirds, most of the
samples were completely used in the unsuccessful creati-
nine assay and hormonal assay. Therefore, to assess the
variation in urine dilution among free-living birds, an
additional study was performed. We recorded SG in the
samples from free-living birds of the three studied species
for which we still had at least 10 lL (N = 43 out of a total
of 94 samples). We compared the SG of these samples with
that of CF from newly captured C. anna that were main-
tained in captivity for 30 min and then released (n = 4).
This study was performed during the reproductive season at
the same dates as the general urine sampling. Birds were
123
J Ornithol
individually maintained in collapsible tulle cages
(60 9 60 9 60 cm); a CF sample was taken when the birds
were caught and then at 15 and 30 min, respectively. All
birds were then released. Ad libitum sucrose solution (25 %
w/w) was offered in a syringe nectar feeder (see below), but
none of the birds fed; therefore, we were able to measure the
effect of dehydration on SG by comparing the SG of the
three CF samples collected during this procedure.
Radioimmunoassay (RIA)
Plasma
After centrifugation, plasma samples (30 lL) were brought
to 200 lL by adding distilled water. Samples were incu-
bated overnight with [3H] steroid (i.e., testosterone or
DHEA, approximately 10,000 cpm, New England Nuclear)
for recovery determination. Steroid extraction was per-
formed with redistilled dichloromethane. Levels of T and
DHEA in plasma were assessed using one direct RIA per
each hormone (e.g., Wingfield et al. 1992). Intra-assay
variation was 8.3 % and 7.9 %, respectively. A detection
limit of 0.19 and 0.20 ng/ml was used for DHEA and T,
respectively.
Cloacal fluid
Prior to analysis of DHEA in cloacal fluid, a sulfatase
enzyme was used to hydrolyze DHEA-S into DHEA (Soma
and Wingfield 2001). For this, 0.75 ml of 0.2 M sodium
acetate buffer (pH 5.0) with sulfatase (Sigma, partially
purified from Helix pomata) was added to 30 lL of each
sample; samples were incubated overnight at 37 °C. The
following day 20 lL of [3H] DHEA was added to each
sample and samples were again incubated overnight.
DHEA extraction was performed with redistilled dichlo-
romethane. RIA was performed as described previously
(Wingfield et al. 1992).
Levels of T in CF were determined using direct RIA
with no extraction (Hiebert et al. 2000). Specifically, the
sample was brought to 200 lL with distilled water; then
100 lL of [3H] T and 100 lL of T antiserum were added
(Hiebert et al. 2000). Inter-assay variation was T 11.74 %.
DHEA samples were run in one assay.
Statistical analysis
To assess the effect of season (reproductive, non-repro-
ductive) on sex steroid levels, we performed a multivariate
analysis of variance (MANOVA). Levels of hormones in
plasma and CF were correlated with a simple regression.
Comparisons between territorial and non-territorial
individuals were assessed utilizing the Mann–Whitney
J Ornithol
test since one of the distributions did not meet the nor-
mality criterion. To assess the effect of DHEA on
aggressive behaviors and body condition, simple regres-
sions were performed. To assess the effect of species on
hormone levels during the reproductive season, an
ANOVA and a posteriori comparison among species
through Tukey HSD for unbalanced sample size was
performed. Differences in T levels between reproductive
and non-reproductive seasons within a species were
compared using the Mann–Whitney test when they did
not meet conditions for normality and the t test when they
did.
To compare SG in CF between free-living and captive
birds at the moment of capture, a Mann–Whitney test was
performed; in addition, a repeated measures Friedman
ANOVA was used to compare SG over time (0, 15 and
30 min) in captive individuals. Analyses were performed
with Statistica 6.0. Data are expressed as mean ± SE.
Results
Laboratory study
Plasma DHEA levels were significantly higher during the
non-reproductive season compared to the early reproductive
season in S. sephanoides (Fig. 1a). In contrast, plasma T was
not significantly different between seasons (Fig. 1a). Steroids
in CF and plasma were positively correlated (Fig. 1b).
Urine dilution
Creatinine in S. sephaniodes was not significantly different
between non-reproductive and reproductive seasons
Fig. 1 Levels of plasma DHEA in S. sephanoides (n = 9) were
significantly higher during the non-reproductive season (black bars)
than during the early reproductive season (Mann–Whitney
U9 = 2.32, P \ 0.05, white bars). In contrast, levels of T were not
(0.30 ± 0.06 and 0.31 ± 0.09 mg/dL, respectively,
mean ± SE; CV = 0.37 %, Mann–Whitney U = 4.5,
P = 0.85). We found no significant relation between cre-
atinine and body condition index (r2 = 0.08; F1,4 = 0.35;
P = 0.58).
Cloacal fluid SG for free-living hummingbirds showed
minimal variation among individuals and species
(1.0037 ± 0.00038, mean ± SE; CV = 0.25 %,
range = 1.002–1.008). There was no difference between
the SG of CF from free-living birds,all three California
species, and from C. anna that was kept in captivity at the
moment of capture (Mann–Whitney U = 78.5, p = 0.72)
or after 15 min of fasting (1.008 ± 0.0019, p = 0.13) but
significantly increased after 30 min (1.011 ± 0.002,
Friedman ANOVA V2 = 6.53, P \ 0.05). On the basis of
these results, we are confident that there is minimal vari-
ation in urine dilution in free-living hummingbirds. We,
therefore, did not incorporate a correction for urine dilution
in the analysis of CF obtained from birds in the field.
Field studies
Males
During the reproductive season, levels of T in CF of A.
alexandri were 3.5 times higher than in Anna’s and 2.6
times higher than in S. rufus (ANOVA (F(1,43) = 3.99,
P \ 0.05, Fig. 2). Levels of T in CF were significantly
greater in A. alexandri than in C. anna (Tukey HSD
P \ 0.05), but not significantly different between A. alex-
andri and S. rufus (Tukey HSD P = 0.15) or between S.
rufus and C. anna (Tukey HSD P = 0.91). Body condition
index was positively correlated with T levels in C. anna
(Fig. 3a), but tended to be negatively correlated with T
significantly different between seasons (U9 = 7.01, P = 0.60, a).
Levels of plasma and urine creatinine referenced DHEA were
positively correlated (r 2 = 0.69, F(1,4) = 8.7, P \ 0.05, n = 5; b)
123

Fig. 2 Testosterone levels in cloacal fluid (CF) during the reproduc-
tive season from males and females of three species of free-living
hummingbirds that were briefly captured to obtain the sample.
Asterisk significantly differences, P \ 0.05
levels in A. alexandri (Fig. 3b). In addition, T levels in CF
of A. alexandri were marginally positively correlated with
the aggression index (Fig. 3c).
During mid-August, in the non-reproductive pre-
migration period, levels of CF T in A. alexandri were 29
times lower than in the reproductive season
(0.79 ± 0.15 ng/ml CF mean ± SE, t = -2.26, P \ 0.05,
n = 17; data not shown); however, the relation between
body mass and hormone levels during this period was not
significant (R2 = 0.22, F(1,8) = 2.4, P = 0.16, n = 10).
During the non-reproductive season, DHEA in CF of
male C. anna was significantly higher in territorial than in
non-territorial males (Fig. 4a). DHEA was positively cor-
related with aggression index (Fig. 4b). Body condition
was marginally positively related to DHEA levels
(Fig. 4c).
Fig. 3 Body condition index was positively related to testosterone in
cloacal fluid (CF) in Anna’s hummingbird (C. anna) during the
reproductive season (r2 = 0.211, F(1,23) = 7.435, P \ 0.05, a).
Although the relation between T and aggression in Black-chinned
123
J Ornithol
Females
Levels of T in CF were not significantly different across
species during the reproductive season (ANOVA
(F(1,21) = 2.78, P = 0.08, n = 24, Fig. 3). T levels in CF
from A. alexandri during the reproductive season
(5.65 ± 1.71 ng/ml) were nine times higher than during
the pre-migration period (0.57 ± 0.19 ng/ml, Mann–
Whitney U = 0.0, P \ 0.05, n = 14; data not shown).
During the reproductive season, we found levels of T in CF
significantly higher in males than in females in C. anna
(Mann–Whitney U = 100, P \ 0.01) and in A. alexandri
(U = 63, P \ 0.05), but not in S. rufus (U = 26,
P = 0.90).
Discussion
In this study, we showed a positive relationship between
sex steroids in plasma and their metabolites in CF
(Fig. 1b). The small blood volumes available from hum-
mingbirds makes this technique especially useful for
studying endocrine mechanisms in these birds. In com-
parison to other birds (Wasser et al. 1993; Goyman 2005),
hummingbirds present advantages to use this methodology
because they have relatively stable diets (López-Calleja
and Bozinovic 2003; González-Gómez et al. unpublished
data) and ingest large amounts of water, which visibly
separates urine from feces.
Reproductive season
As it was earlier reported by Stiles (1971), C. anna did not
defend feeding territories consistently during reproductive
season, although they displayed inter-specific and intra-
(A. alexandri) was marginally positive during the reproductive season
(r2 = 0.75, F(1,3) = 8.13, P = 0.058, c), we found that body condi-
tion tended to be lower in birds with higher T (r2 = 0.55, F(1,5) = 6.1,
P = 0.055, b)
J Ornithol
Fig. 4 In cloacal fluid (CF) obtained during the non-reproductive
season from free-living C. anna, DHEA was significantly higher in
territorial males than in non-territorial males (Mann–Whitney test
U = 2.00, P \ 0.05, n = 9, a). During the non-reproductive season,
specific aggressiveness. Therefore, we were unable to
relate territorial aggression levels with sex steroid levels;
we did, however, observe that higher levels of T were
positively related with body condition (Fig. 3a). This
finding supports the idea that aggression levels may be
related to resource acquisition and could provide an honest
signal for sexual selection, either directly, in displays of
physical attractiveness or motor performance, or indirectly,
through the acquisition and defense of high quality food
sources (McGlothlin et al. 2007; Byers et al. 2011).
When A. alexandri arrived at our field site in early April,
they were consistently territorial and actively displaced C.
anna from feeders. We found aggression levels in this
species tended to be positively correlated with T (Fig. 3b).
However, we observed a marginally negative relation
between T and body condition (Fig. 3c). Possible expla-
nations could be related to the lower capacity of A. alex-
andri for fat storage in comparison to bigger hummingbird
species such as C. anna (Hainsworth 1978). In addition, A.
alexandri males face higher energy expenditures than C.
anna during the reproductive season. They perform a long
distance migration to breeding grounds where they defend
feeding resources at the same time that they perform
expensive courtship displays. In middle August, during
molt and before migration, A. alexandri showed T levels
significantly lower in comparison to the reproductive sea-
son. This was expected at the end of the breeding season as
it does in many other seasonal species (Wingfield et al.
1990, 2001). In addition, elevated levels of T could inhibit
or delay molt (Schleussner et al. 1985).
Conversely, we found high variation in T levels in S.
rufus, which could be at least partially explained by the
location of our field site. Specifically, our field site is in the
migration route of this species. Therefore, it is possible
some of the birds could have been summer residents while
others were likely to be migrants heading north. Further
research is needed to assess the link between survival,
aggression was positively correlated with the DHEA levels in CF
(F(1,7) = 6.29, P \ 0.05, n = 9, b). In addition, the body condition
index was marginally positively related with the hormone level in C.
anna (F(1,7) = 4.98, P = 0.06, n = 9, c)
androgen levels, reproductive success, and energy expen-
diture in long distance migrant hummingbirds.
Non-reproductive season
In S. sephanoides, plasma levels of DHEA were higher
than T during the non-reproductive season (Fig. 1). How-
ever, during the (early) reproductive season, we found
plasma levels of DHEA to be lower and levels of T highly
variable. We found a consistent pattern in C. anna where
levels of DHEA were significantly higher in territorial than
in non-territorial individuals and positively correlated with
aggression levels. In this species, body condition tended to
be higher in males with higher DHEA (Fig. 4c), suggesting
that more aggressive subjects were able to secure higher
quality resources (Copenhaver and Ewald 1980; Camfield
2006). Our results could suggest that androgen precursors
such as DHEA are linked to aggression levels during the
non-reproductive season in year-round territorial hum-
mingbird species; however new studies and manipulative
experiments are needed to confirm this hypothesis.
In our field site, the aggression displayed by C. anna
(measured with the aggression index) was five times lower
than that previously reported for S. sephanoides during the
same time of the year (González-Gómez et al. 2011a).
These behaviors were displayed to maintain feeding terri-
tories during the non-reproductive season (Stiles 1971) and
included chases, contacts, and vocalizations. Despite cli-
matic conditions and habitat structure similarities in the
areas inhabited by C. anna in central California and S.
sephanoides in central Chile, C. anna maintained territories
that were roughly five times bigger than typical S. sep-
hanoides territories (Stiles 1971; González-Gómez and
Vasquez 2006). This may offer an explanation at the eco-
logical level for the observed differences in aggression
levels during the non-reproductive season. Previous studies
have shown that under similar resource conditions, the
123

intensity of territory defense tends to be higher in smaller
territories and where open vegetation structure easily allows
territory owners to detect intruders (Norton et al. 1982).
Female aggression
Although we did not quantify aggression in females, we did
observe intra-specific and inter-specific aggressive
encounters among female C. anna, A. alexandri, and S.
rufus. Even when females did not display territorial
behaviors consistently, levels of T in CF of female S. rufus
were similar to those in males of this species. Previous
studies demonstrate that levels of T in females of sexually
polygynous species are significantly lower than levels of T
in conspecific males and lower than levels of T in females
of monogamous species (Ketterson et al. 2005). Because
hummingbirds are sexually polygynous, we expected T in
males and females to show a similar pattern; the fact that
they did not could be an effect of small sample size and/or
large variation in T levels of males.
The fact that females show detectable levels of T during
the reproductive season may be related to the role of T in
ovulation and egg production. For example, during ovulation
the thecal layer of the developing follicle produces T, which
is converted to an estrogen by adjacent granulosa cells
(Gomez et al. 2001). T also regulates the production of
albumin by the oviduct (Yu and Marquardt 1973; Müller
et al. 2002). In addition, T that is not converted to estradiol in
the ovary can circulate and affect behavior (Brenowitz and
Arnold 1993). Considering that female hummingbird species
solely display parental care, it would be very difficult for
females to hold a feeding territory during the reproductive
season. However, aggressive behaviors against other hum-
mingbird males or females could be advantageous in terms of
improving food access (González-Gómez et al. 2011a). That
could be especially significant in species such as A. alexandri
and S. rufus, in which males are highly territorial and
aggressive during this period, often forcing the females to
poor and scattered resources (Kodric-Brown and Brown
1978). On the other hand, it has been argued that aggression
in females could be related to circulating levels of estradiol
(Pärn et al. 2008, but see Elekonich and Wingfield 2000). Our
preliminary data show that at the end of the reproductive
season, estradiol in females was 25 times higher than levels
of T, but further sampling would be necessary to address the
relationship between estradiol and aggressive behaviors in
female hummingbirds.
Conclusion
High levels of steroids supporting territorial behaviors are
likely to be adaptive in territorial animals. Our results
123
J Ornithol
suggest that the regulation of aggressive behavior in
hummingbirds and, therefore, the ability to secure scarce
resources, could be mediated by different hormones during
the reproductive season and non-reproductive seasons in
year-round territorial species such as S. sephanoides and C.
anna. Manipulations of DHEA and T are needed to confirm
the functional relationship between hormone levels,
aggressive behaviors and body condition. Further research
is needed to describe these relationships through the annual
life cycle, especially in migratory species.
Acknowledgments This paper is dedicated to the memory of Loreto
Godoy. We thank Baggins End Community, and Lee James of Tierra
Verde Vegetables for allowing us to work in their properties. We
thank Kara Sweeney and Holly Ernest for contacts and sampling help.
This research was supported by CONICYT, Postdoc Becas Chile to
PGG and the Endowed Professorship in Physiology to JCW, College
of Biological Sciences, University of California, Davis and Grant
Number IOS-0750540 from the National Science. Foundation to
JCW.
References
Armstrong LE (2005) Hydration assessment techniques. Nutr Rev
63:S40–S54
Brenowitz EA, Arnold AP (1993) Hormone accumulation in song
regions of the canary brain. J Neurobiol 23:871–880
Byers J, Hebets E, Podos J (2011) Female mate choice based upon
male motorperformance. An Behav 79:771–778
Camfield AF (2006) Resource value affects territorial defense by
broad tailed and rufous hummingbirds. J Field Ornithol
77:120–125
Copenhaver C, Ewald PW (1980) Cost of territory establishment in
hummingbirds. Oecologia 46:155–160
Elekonich MM, Wingfield JC (2000) Seasonality and hormonal
control of territorial aggression in female song sparrows
(Passeriformes: Emberizidae: Melospiza melodia). Ethology
106:493–510
Gomez Y, Velazquez PN, Peralta-Delgado I, Mendez MC, Vilchis F,
Juarez-Oropeza M, Pedernera E (2001) Follicle stimulating
hormone regulates steroidogenic enzymes in cultured cells of the
chick embryo ovary. Gen Comp Endoc 121:305–315
González-Gómez PL, Estades CF (2009) Is natural selection
promoting sexual dimorphism in the Green-backed firecrown
hummingbird (Sephanoides sephaniodes)? J Ornithol
150:351–356
González-Gómez PL, Vasquez R (2006) A field study of spatial
memory in Sephanoides sephaniodes. Ethology 112:790–795
González-Gomez PL, Ricote-Martı́nez N, Razeto-Barry P, Bozinovic
F (2011) Thermoregulatory cost affects territorial behavior in
hummingbirds: a model and its application. Behav Ecol Socio-
biol 65:2141–2148
González-Gómez PL, Bozinovic F, Vásquez RA (2011a) Elements of
episodic-like memory in free-living hummingbirds, energetic
consequences. An Behav 81:1257–1262
González-Gómez PL, Vásquez RA, Bozinovic F (2011b) Flexibility
of foraging behavior in hummingbirds: the role of energy
constraints and cognitive abilities. Auk 128:36–42
Goyman W (2005) Noninvasive monitoring of hormones in bird
droppings physiological validation, sampling, extraction, sex
J Ornithol
differences, and the influence of diet on hormone metabolite
levels. Ann NY Acad Sci 1046:35–53
Goymann W, Landys MM, Wingfield JC (2007) Distinguishing
seasonal androgen responses from male–male androgen respon-
siveness—revisiting the challenge hypothesis. Horm Behav
51:463–476
Hainsworth FR (1978) Feeding: models of costs and benefits in
energy regulation. Am Zool 18:701–714
Hau M (2007) Regulation of male traits by testosterone: implications
for the evolution of vertebrate life histories. Bioessays
29:133–144
Hau M, Wikelski M, Soma KK, Wingfield JC (2000) Testosterone
and year-round territoriality in a tropical bird. Gen Comp Endoc
117:20–33
Hiebert SM, Ramenofsky M, Salvante K, Wingfield JC, Gass CL
(2000) Noninvasive methods for measuring and manipulating
corticosterone in hummingbirds. Gen Comp Endoc 120:235–247
Ketterson ED, Nolan V Jr, Sandell M (2005) Testosterone in females:
mediator of adaptive traits, constraint on the evolution of sexual
dimorphism, or both? Am Nat 166:S85–S98
Kodric-Brown A, Brown JH (1978) Influence of economics, inter-
specific competition, and sexual dimorphism on territoriality of
migrant rufous hummingbirds. Ecology 59:285–296
López-Calleja MV, Bozinovic F (2003) Dynamic energy and time
budgets in hummingbirds: a study in Sephanoides sephaniodes.
J Comp Biochem Physiol A 134:283–295
McGlothlin JW, Jawor JM, Ketterson ED (2007) Natural variation in
a testosterone-mediated trade-off between mating effort and
parental effort. Am Nat 170:864–875
Müller W, Eising C, Dijkstra C, Groothuis T, Sageder G (2002) Sex
differences in yolk hormones depend on maternal social status in
leghorn chickens (Gallus gallus domesticus). Proc R Soc Lon B
269:2249–2255
Norton ME, Arcese P, Ewald PW (1982) Effect of intrusion pressure
on territory size in black-chinned hummingbirds (Archilochus-
alexandri). Auk 99:761–764
Ogawa S, Lubahn D, Korach K, Pfaff D (1997) Behavioral effects of
oestrogen receptor gene distribution in male mice. Proc Natl
Acad Sci USA 94:1476–1481
Pärn H, Lindström KM, Sandell M, Amundsen T (2008) Female
aggressive response and hormonal correlates—an intrusion
experiment in a free-living passerine. Behav Ecol Sociobiol
62:1665–1677
Schleussner G, Dittami JP, Gwinner E (1985) Testosterone implants
affect molt in male European Starlings, Sturnus vulgaris. Physiol
Zool 58:597–604
Soma KK, Wingfield JC (2001) Dehydroepiandrosterone in song
plasma: seasonal regulation and relationship to territorial
aggression. Gen Comp Endoc 123:144–155
Soma KK, Traimontin AD, Wingfield JC (2000) Oestrogen regulates
male aggression in the non-breeding season. Proc Roy Soc Lon
B 267:1089–1096
Soma KK, Wissman AM, Brenowitz EA, Wingfield JC (2002)
Dehydroepiandrosterone (DHEA) increases territorial song and
the size of an associated brain region in a male songbird. Horm
Behav 41:203–212
Stiles FG (1971) Time, energy, and territoriality of the anna
hummingbird (Calypte anna). Science 173:818–820
Suarez RK (1992) Hummingbird flight: sustaining the highest mass-
specific metabolic rates among vertebrates. Experientia
48:565–570
Suarez RK, Gass CL (2002) Hummingbird foraging and the relation
between bioenergetics and behaviour. Comp Biochem Physiol A
133:335–343
Temeles EJ, Goldman RS, Kudla AU (2005) Foraging and territory
economics of sexually dimorphic purple-throated caribs (Eu-
lampis jugularis) on three heliconia morphs. Auk 122:187–204
Wasser SK, Thomas R, Nair PP, Guidry C, Southers J, Lucas J, Wildt
DE, Monfort SL (1993) Effects of dietary fiber on faecal steroid
measurements in baboons (Papio cynocephalus cynocephalus).
J Rep Fertil 97:569–574
Wingfield JC (1984) Androgens and mating systems: testosterone-
induced polygyny in normally monogamous birds. Auk
101:665–671
Wingfield JC, Soma KK (2002) Spring and autumn in song sparrows:
same behavior, different mechanisms? Integr Comp Biol
42:11–20
Wingfield JC, Henger RE, Dufty AM Jr, Ball GF (1990) The
‘‘challenge hypothesis’’: theoretical implications for patterns of
testosterone secretion, mating systems, and breeding strategies.
Am Nat 136:829–846
Wingfield JC, Vleck CM, Moore MC (1992) Seasonal changes of the
adrenocortical response to stress in birds of the sonoran desert.
J Exp Zool 264:419–428
Wingfield JC, Lynn S, Soma KK (2001) Avoiding the ‘costs’ of
testosterone: ecological bases of hormone–behavior interactions.
Brain Beh Evol 57:239–251
Yu JYL, Marquardt RR (1973) Synergism of testosterone and
estradiol in the development and function of the magnum from
the immature chicken oviduct. Endocrinology 92:563–572
123